On Tue, 2022-06-28 at 23:10 +0700, Ammar Faizi wrote:
> On 6/28/22 10:04 PM, Dylan Yudaken wrote:
> > copy from include/uapi/linux/io_uring.h
> >
> > Signed-off-by: Dylan Yudaken <dylany@xxxxxx>
> > ---
> > src/include/liburing/io_uring.h | 53 ++++++++++++++++++++++++----
> > -----
> > 1 file changed, 39 insertions(+), 14 deletions(-)
> >
> > diff --git a/src/include/liburing/io_uring.h
> > b/src/include/liburing/io_uring.h
> > index 2f391c9..1e5bdb3 100644
> > --- a/src/include/liburing/io_uring.h
> > +++ b/src/include/liburing/io_uring.h
> > @@ -10,10 +10,7 @@
> >
> > #include <linux/fs.h>
> > #include <linux/types.h>
> > -
> > -#ifdef __cplusplus
> > -extern "C" {
> > -#endif
>
> Dylan,
>
> That `extern "C"` thing is for C++, we shouldn't omit it.
>
> Or better add that to the kernel tree as well, it won't break
> the kernel because we have a __cplusplus guard here.
>
> Jens what do you think?
>
> Just for reference, that line is introduced in commit:
>
> commit 3d74c677c45eccf36b92f7ad4b3317adc1ed06bb
> Author: Bart Van Assche <bvanassche@xxxxxxx>
> Date: Sun Jun 28 12:58:19 2020 -0700
>
> Make the liburing header files again compatible with C++
>
> Include <atomic> instead of <stdatomic.h> if built with a C++
> compiler.
>
> Fixes: b9c0bf79aa87 ("src/include/liburing/barrier.h: Use C11
> atomics")
> Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
>
> src/include/liburing.h | 8 ++++----
> src/include/liburing/barrier.h | 37
> +++++++++++++++++++++++++++++++++++--
> src/include/liburing/io_uring.h | 8 ++++++++
> 3 files changed, 47 insertions(+), 6 deletions(-)
>
> Adding the author to the CC list.
>
Yeah this makes sense. I think I just assumed the file was a manual
copy of the latest kernel one which it clearly is not.
I'll fix it up in v2
